Introduction

Transient Elastography (TE) is a well-validated, non-invasive method to assess hepatic fibrosis and steatosis using liver stiffness measurement (LSM) and controlled attenuation parameter (CAP), respectively. However, due to interobserver variability, uncertainties still exist regarding the reliability and reproducibility of the technique especially in obese patients.

Aims & Methods

We aimed to evaluate the interobserver agreement of transient elastography in overweight and obese patients with Nonalcoholic Fatty Liver Disease (NAFLD).
We conducted a cross-sectional study including obese patients with NAFLD. We analyzed the results of TE performed by two different experienced operators, on two different days, within the same month. For each patient, the probe size (M or XL) used in both exams was the same; the weight remained stable and all patients had normal liver function. LSM was recorded after obtaining at least 10 valid readings, a success rate ≥ 60% and an interquartile range to a median ratio (IQR/M) < 0.3. We classified the risk for advanced fibrosis as low, indeterminate and high, based on EASL guidelines 2021, which were respectively: <8 kPa, ≥8 and < 12 kPa, and ≥12 kPa. Hepatic steatosis was defined by CAP values ≥ 248 dB/m.

Results

Sixty-eight patients were recruited (70.6% female) with a mean age of 53.21±12.9 years. The mean body mass index (BMI) was 42.05±9.57 kg/m2 and the mean waist circumference (WC) was 119.51±22.07 cm. 10.4% of patients were overweight, 18.8% were obese and 70.8% were severely obese. There was no statistically significant difference in both mean LSM (7.3 vs 6.9 kPa; P=0.475) and mean CAP values (313 vs 309.8 dB/m; P=0.507), assessed by the first and the second operators, respectively. The measurements made by operators of LSM and CAP were correlated [(Spearman'sρ = 0.512; P< 0.001) and (Spearman'sρ = 0.525; P< 0.001), respectively]. The interobserver intraclass correlation coefficient was 0.860 (95% CI 0.772-0.914) for LSM and 0.710 (95% CI 0.526-0.822) for CAP values. A discrepancy of at least 2 kPa between both results was noted in 30 cases (44.1%). There was no statistically significant difference in mean IQR/M values between the two exams (0.14 vs 0.12; P=0.15). The discordance between operators was 16.2% for at least one level of risk of advanced fibrosis and 14.7% for two levels. In fact, among the patients considered with a high risk of advanced fibrosis by the first operator, 60% had low risk, 20% had indeterminate risk and 20% had high risk, assessed by the second one. For those who had an indeterminate risk of fibrosis, according to the measurements made by the first operator, 60% of them were at low risk of advanced fibrosis, based on those made by the second one. Among those at low risk of advanced fibrosis assessed by the first operator, 8.7% had a high risk of advanced fibrosis as measured by the second one. Interobserver reliability values were κ = 0.32 for low risk of advanced fibrosis, κ = 0.12 for indeterminate risk, and κ = 0.16 for high risk for advanced fibrosis. The interobserver discrepancy in LSM was significantly associated with LSM (P=0.001), BMI (P<0.001), and WC (P=0.025).

Conclusion

Although TE is a very convenient method to assess liver fibrosis and steatosis in clinical practice, the interobserver discrepancy in results was not negligible, especially in severely obese patients and mainly for indeterminate and high risk of advanced fibrosis, which represents a significant problem with the technique. Thus, the evaluation of liver fibrosis, including TE results, should be carefully analyzed by expert hands.